From: Jamal Mazrui <[EMAIL PROTECTED]>
>From the web page http://www.nfb.org/csu99win.htm
COMPUTER SCIENCE UPDATE--WINTER-SPRING, 1999
Published By
National Federation of the Blind in Computer Science
President, Curtis Chong
1800 Johnson Street
Baltimore, MD 21230
Phone: (410) 659-9314
E-mail: [EMAIL PROTECTED]
Developing a Windows Screen Reader
>From the Perspective of GW Micro
by Doug Geoffray
>From the editor: At the 1998 meeting of the National Federation
of the Blind in Computer Science held in Dallas, Texas, during
the first week of July, a good chunk of time was devoted to
presentations from two screen access vendors on the subject of
developing a screen reader for Windows: Henter-Joyce and Artic
Technologies; GW Micro was invited to participate but was unable
to do so for reasons that will become apparent later in this
article. As Microsoft works continually to improve its operating
systems and application programs, screen access vendors are of
necessity forced to keep up or lose their place in the market.
In other words, whatever Microsoft does to its software, screen
access vendors are affected--sometimes to the good and sometimes
not. The presentations by Henter-Joyce and Artic Technologies
International underscored this fact. Since GW Micro has one of
the leading screen reading programs for Windows on the market,
and since I had invited Doug Geoffray to speak on behalf of the
company, I thought it would be appropriate and even necessary
for the company's point of view to be presented here.
As you will see, GW Micro has taken the position that it will
continue to support Microsoft's Active Accessibility programming
interface, despite many problems and frustrations. As of this
writing, I can tell you that the wisdom of this decision is now
beginning to be revealed. Very recent tests conducted at the
Federation's International Braille and Technology Center for the
Blind with GW Micro's Window-Eyes Version 3.0 and Microsoft's
Internet Explorer Version 5 are very encouraging. What I will
say at this point is that with both programs taking advantage of
the latest version of Active Accessibility, our ability to get
at information on hitherto difficult-to-read web pages improved
dramatically.
Here, then, is what Doug Geoffray of GW Micro has to say:
I wish to apologize for not being able to personally attend this
meeting. But given the choice of attending this meeting or
getting married, well, let's say the choice was easy. Sorry
Curtis but I don't have to spend the rest of my life with you.
Okay, on to business.
I must preface this by saying I have signed a non-disclosure
agreement with Microsoft. There are certain things which I can
not legally discuss. I have already gotten my hands slapped a
couple of times because of this.
We all know that applications written using standard Windows
controls and conventions will be completely accessible. For
example, if the application uses standard edit boxes and list
boxes and gives the user keyboard access for all options, it is
going to work out of the box with Window-Eyes. However, because
competition is so intense for Windows applications, developers
are confronted with development decisions. Many of those
decisions deal with how information is presented to and
retrieved from the end user.
A standard edit box just doesn't have enough pizzazz anymore.
How about a snazzy super-duper edit control? How about writing
information vertically instead of horizontally? That will surely
impress potential users.
Obviously, applications can get much more involved, and this is
what causes the problem. How can adaptive applications like
Window-Eyes get into these applications with high appeal and
make them usable by a blind person?
The fact is that Window-Eyes can't possibly be modified for each
and every application. Something else needs to be done.
Microsoft recognized the problems adaptive vendors were facing.
Microsoft and adaptive vendors got together and hashed out an
interface (Active Accessibility) which applications could choose
to use. By choosing to use this interface, adaptive applications
wouldn't need to guess what was happening. The application could
just tell the adaptive software directly and/or the adaptive
software could just ask the application. Wow, what potential!
The problem is how to create an interface specific enough to get
the job done but generic enough to encompass the wide variety of
application programs.
There was much debate over what this technology should and
should not do. One of the options was whether the interface
should incorporate an off-screen-model (OSM). Every screen
reader needs to have an exact image of what is displayed on the
screen. In the DOS days, the screen reader could simply read the
information directly from the video card at any time. However,
with Windows, this is not possible. There needs to be a database
off screen which represents what is displayed.
Microsoft, in my opinion, made a token gesture by purchasing an
existing adaptive company's OSM. I say gesture simply because
there was very little research done before purchasing the OSM.
Microsoft knew it needed to do something quick to settle people
down. The purchase was made for an undisclosed amount of money
but never used. Aside from giving the developer some extra cash
and temporary marketing hype, nothing further happened. The OSM
technology was never used and never will be.
We at GW Micro believed in the potential of Microsoft Active
Accessibility (MSAA) from the beginning but quickly became
frustrated with Microsoft. The communication between our two
companies was very sporadic. The beta copies of MSAA were
plagued with bugs making it impossible to use or test. This went
on for quite a long time. We began to wonder if MSAA was really
going to be a reality. So we opted to leave it alone for awhile.
Towards the middle of 1997, we took another serious look at MSAA
and were pleased to see a somewhat bug-free version. We went
full speed ahead with integrating MSAA into Window-Eyes. At the
time, the only applications supporting MSAA were Office 97 and
Internet Explorer 3.0x. Although these are two huge programs,
obviously more applications would have to support MSAA, and we
gambled that more would be coming.
Microsoft released MSAA 1.0 to the public in May of 1997.
Although this version was lacking in some features, it was
relatively bug free. It was a good stepping stone into better
access. The Window-Eyes development continued with a deep
integration of MSAA. But then Microsoft made a couple of huge
blunders which many of you are already aware of.
One of the biggest blunders was releasing Internet Explorer 4.0
with no support for accessibility. Microsoft has been tooting
their horn about MSAA and how it is the future but then released
Internet Explorer (IE) 4.0 without it. This was a tremendous
setback for Microsoft and any credibility it may have had. This
demonstrates the incredible pressure placed on software
manufacturers. Get that product out on the market. We can't deal
with accessibility issues or we'll loose our market advantage.
Somehow this mentality must change. It would be much more
desirable if corporations say to themselves, "If we don't get
accessibility added to our application, we will loose our market
advantage." I honestly see this day coming, but it won't be an
easy road.
So, feeling the pressure and potential boycotts, Microsoft
released IE 4.01 about a month later. This version was supposed
to add the necessary accessibility components allowing
applications like Window-Eyes full support. I consider this
blunder number two. As it turned out, IE 4.01 was just a token
effort toward accessibility. The communication between Microsoft
and GW Micro was all but nonexistent.
Along with the release of IE 4.01 Microsoft released MSAA 1.1,
which was required for IE 4.01. Microsoft never told us that MSAA
1.1 was going to be released. It just happened. Microsoft was
telling everyone that IE 4.01 and MSAA 1.1 were well tested and
worked with adaptive applications. Well, I'm not sure what
software was used to conduct the testing, but there was no
screen reader for the blind on the market that supported IE 4.01
and MSAA
1.1. The latter was and still is plagued with bugs.
Consequently, we at GW Micro decided to stay with MSAA 1.0.
By this time Microsoft was taking on a lot of well-deserved heat
about these blunders. Was Microsoft truly interested in making
its applications accessible? There was a lot of lip service but
nothing to show for it. Vendors like ourselves were beginning to
wonder if we should become dependent on MSAA if Microsoft was
going to keep blundering. We certainly didn't want to make
Window-Eyes dependent on a technology that apparently had no
clear future. It was a difficult decision, but we opted
ultimately to stick with MSAA and Microsoft.
In February 1998, Microsoft put on an accessibility day for all
of its employees. Representatives from a number of adaptive
technology companies (including GW Micro) and several advocate
representatives were flown to the company headquarters in
Redmond, Washington. This allowed all Microsoft employees the
opportunity to become aware of the existing technology available
for persons with disabilities to use Microsoft programs.
In some sense, the day was a large success. It allowed me to
meet face to face with many of the project leaders for many of
Microsoft's top applications. Being able to express the problems
directly with them was very helpful. Interestingly enough, it
became clear that Microsoft had not planned for adaptive
technology vendors and consumer advocates to meet with each
other. Only after the advocates demanded it was a meeting
scheduled. Did Microsoft deliberately plan on keeping the two
groups apart? Certain plans were introduced to the advocates
which the vendors never heard. Why was Microsoft doing this? Why
couldn't communication between all interested parties have been
improved and made consistent?
While attending the accessibility day, I was asked by several
advocates what direction they should be taking with Microsoft.
They wanted to know exactly what they should be advocating for
to help us the most.
One of the issues brought up was whether Microsoft should
develop its own off-screen model. The question was asked, if
Microsoft did create such a model, would GW Micro utilize it?
Because of the existing investment we have in our own off-screen
model, I don't believe we would. We have modeled Window-Eyes
based on the abilities of our own off-screen model. It would
involve a large undertaking to modify the entire interface to
use the Microsoft off-screen model.
One argument says, "Don't you think Microsoft could create the
best off-screen model?" Based on the past performance, I don't
believe Microsoft could do a better job then we have already
done. We also have the problem of being dependent on Microsoft
to make enhancements and fix bugs in a timely manner. Again,
judging from its past performance, I don't believe this would be
possible.
Another argument supporting Microsoft's creation and
implementation of its own off-screen model is that more
companies would produce screen readers. Although this may be
true, there is much more to a screen reader than the off-screen
model. I think we would indeed see more screen readers hit the
market, but their quality would not and could not compete with
the screen readers of today.
Another issue which is heating up is this: should Microsoft
create its own large print and screen reader applications? At
first, everyone might say "of course." But when you think about
it, it may have a negative effect on blind people.
Microsoft will be producing a very basic screen reader and large
print application that will ship with its operating systems. It
claims it is doing this so blind and visually impaired people
can install operating systems, upgrade operating systems, and
have access to the control panel. These utilities will be very
basic in nature. They will have no features and would not be
useable for more complicated applications. I believe this is
great. However, there is much concern that Microsoft will
continue to enhance these utilities over time.
Again, at first, you as a blind person may be saying "what's the
problem with that?" My response would be competition. If
Microsoft continues to evolve these utilities and give them away
with the operating system, the competition will die out. How can
you compete with a free screen reader? Again, you may be saying
"so what, I would rather use the free screen reader." That may
be true but do you really want Microsoft to be the only game in
town? It would ultimately dictate what the utilities should and
should not do. Do you want to deal with Microsoft for technical
support issues? This can go on and on. The bottom line:
Microsoft should supply very basic utilities to guarantee that
an individual can install, upgrade, and configure the operating
system but should stop at that.
I personally would rather see Microsoft take its resources and
get the bugs out of MSAA. Enhance the interface to MSAA. Add
MSAA to all (and I mean all) Microsoft applications. Market MSAA
to other software manufacturers. The company can't possibly do
this and create an off-screen model, full-featured screen
enlarger, and full-featured screen reader. I believe the biggest
bang for the buck is to let Microsoft work on the interface
technology and let companies like GW Micro create the actual
utilities.
The accessibility group at Microsoft is always asking how it can
do a better job. My first response is always better
communication. I must admit that communication has gotten better
over the years. But there are still times when we, the access
vendors, are left totally in the dark. If Microsoft is truly
interested in making its products accessible, companies like GW
Micro must be deeply involved. Once the product is released, it
is too late. Three perfect examples of this are IE 4.0, IE 4.01,
and MSAA 1.1.
So my experience with Microsoft has definitely had its ups and
downs. There are times when it seems to care less about
accessibility vendors and other times when it seems deeply
committed. But GW Micro still supports Microsoft 100% when it
comes to MSAA. We have proven this technology can and does work
with our Window-Eyes 2.1 release. This release offers tremendous
access to the Internet using IE 3.0x. It's the MSAA support in
IE 3.0x which gives us this ability.
We are working hand in hand with Microsoft in many other areas
including IE 5. I believe IE 5 will work as good if not better
than IE 3.0x. We are also working with Microsoft on IE 4.01. So
things are looking brighter. It has been a very difficult
journey and will continue to be difficult for awhile longer. But
in the end, good things are coming.
I also wanted to mention that I believe it is important that GW
Micro continue to back MSAA. Other vendors have decided not to
implement MSAA for some of the reasons I discussed. But in the
end I believe this is a huge disservice to the community.
Without backing from adaptive vendors, Microsoft has no chance
of getting MSAA working and incorporated in major applications.
In the short term, companies that don't support MSAA may look to
be doing better but in the end, it will collapse. MSAA is the
future.
Before I finish, I should clarify one thing. I wholeheartedly
believe the entire accessibility team at Microsoft is committed
to making all products accessible. I have never seen one bit of
evidence to support any other conclusion. I have seen many
attacks on Microsoft and many are directed at the accessibility
team when most attacks should be more appropriately aimed at
Microsoft in general. I would suggest you use the accessibility
team as a funnel to get your thoughts and options through to
Microsoft, but they shouldn't necessarily be directed at the
accessibility team. This team is our only hope. Let's try and
work with them and not against them.
----------
Summary of Remarks by Ted Henter
President Henter-Joyce, Inc.
>From the editor: Here is a summary of the remarks presented by
Ted Henter at the 1998 meeting of the National Federation of the
Blind in Computer Science. The summary was supplied to me by
Henter-Joyce. Contrast this summary with the remarks of Greg
Lowney (see previous article). As a point of information,
readers should be aware that Dale McDaniel, who presented on
behalf of Artic Technologies at the meeting, expressed agreement
with what Ted Henter said.
The following summary raises some interesting questions. What
leverage do we really have over Microsoft to force greater
accessibility to its software? How far should Microsoft go in
terms of developing its own screen magnifier and talking
programs for Windows? More to the point, how much will the
proposed Magnifier (which already comes with Windows 98) and
Narrator (which is slated to be delivered with Windows 2000)
hurt the access market? These are important questions about
which there is much speculation. Certainly, we as consumers do
not want to see the screen reader market abolished. However, we
also want Microsoft to deliver on the commitments it has made.
Here, then, is a summary of Ted's remarks:
Microsoft continues to develop and release products with little
or no regard for compatibility with existing Access products,
like screen readers and magnifiers. As recently as the Windows
98 release, no testing was done with JAWS, and probably the
other access tools. They create problems that we have to fix.
They (Microsoft) are developing the underlying pieces of
software that allegedly will make access easier, even though
these have already been developed by the current access
developers. They have stated that the goal is to "level the
playing field", i.e. reduce the success of the current leaders
in the Access industry by helping others, and themselves, to
develop the same solutions. Traditional access developers will
not be able to continue their current high level of research and
development, for economic reasons. Why develop it yourself when
Microsoft provides it for free?
When they develop the code, we will all be dependent on them,
and they will control the access.
Why would they want to do this? When Microsoft develops the
access tools, then Microsoft controls the access, and then only
Microsoft products will be accessible. Then only Microsoft
products will be bought by the Government. Federal, State, and
Local government procurements will be limited to Microsoft
products only. The Government is the single biggest buyer of
computer software, amounting to billions of dollars each year.
This is the reason why Microsoft wants to get into and control
the access business.
Why is this bad for blind people? Only Microsoft products will
be accessible. Say goodbye to WordPerfect, Netscape, Lotus,
Quicken, Novel, Oracle, 3270 emulators, Eudora, Borland, etc.
But just as bad, the level or quality of access will be governed
by Microsoft, who will have no incentive to make it better. They
need only convince the Government that it is acceptable. They
will not be driven by the competitive forces that have made the
current level of access as good as it is, in spite of the level
of cooperation from Microsoft. Current Access Developers strive
to beat the competition and make the consumer happy. Microsoft
will be concerned about neither. They will have no competition,
and the only consumer they want is the Government.
Including a low quality magnifier and screen reader as part of
the operating system will create confusion among purchasers.
These products will get better and better as time goes on,
further adding to the confusion, preventing the purchase and use
of the right high quality tools even when such tools are needed
by the user. Diminished sales will hinder the ability of the
current access developers to develop the high quality solutions,
and they will cease to exist. The overall level of access will
go down, and our choices will disappear.
What can we do about it? Convince the other mainstream
developers that there is good reason to have access, like all
the Government dollars. They do not have to rely on Microsoft to
provide the solutions, which would definitely be unpalatable for
some of Microsoft's competitors. They should work with the
independent access developers to make their products accessible.
Convince Microsoft that building Access products into their
operating systems is not in the best interest of blind computer
users and that they should not do it. Encourage them to work
with independent developers to solve the problems we face every
day instead of re-inventing the same solutions and techniques.
They should work towards solving the problems, not monopolizing
the market.
----------
End of Document
-
Visit the jfw ml web page: http://yoyo.cc.monash.edu.au/~nallan/jfw
